Research On Distribution Network Reconfiguration Based On EDAPO

Distribution network plays important role in power system.Losses of distribution network are the most among losses of transmission allocation.Because of load imbalance and existence of non-linear load,economic problems of distribution network have been attracting more and more attention.Due to serious lOkV distribution network losses,optimizing the network structure is an important means to improve the economy and reduce network losses.According to objective function and constraints,this paper introduces Artificial Physics Optimization(APO)and Estimation of Distribution Algorithms(EDAs),and combined organic composition of the two new optimization algorithm to apply to lOkV distribution network reconfiguration.According to the characteristics of the closed-loop distribution network structure and open-loop operation when distribution network operates normally,this paper is based on the node degree topological analysis of the distribution network fast,and selection is used to the adjacency list of ways to store sparse data distribution network.In power flow calculations,it uses hierarchical forward and backward extrapolation algorithms to improve the algorithm's parallelism,and through the power flow solution we establish this paper's fitness function,that is,the minimum network loss.We base on the size of the fitness function values to conduct network reconfiguration.This paper introduces the Artificial Physics Optimization(APO)and Estimation of Distribution Algorithms(EDAs).We take advantage of complementarities to combine and get a new organic hybrid algorithm(EDAPO),thereby to increase the algorithm's global search and local search,make the new algorithm converge faster convergence to the global optimum.This paper applies a new hybrid optimization algorithm to network reconfiguration.It utilizes the node degree method to analyze distribution network frame topology.Concentrated coding scheme is used during distribution network coding,which reduces calculation workload largely.Finally,we use Matlab platform programming,through the IEEE33 node and the U.S.P&G69 node system simulation to verify that the algorithm is reasonable and effective.